Chapter 12 Support for ObjectOriented Programming ISBN 0

Chapter 12 Support for Object-Oriented Programming ISBN 0 -321 -33025 -0

Chapter 12 Topics • • • Introduction Object-Oriented Programming Design Issues for Object-Oriented Languages Support for Object-Oriented Programming in Smalltalk Support for Object-Oriented Programming in C++ Support for Object-Oriented Programming in Java Support for Object-Oriented Programming in C# Support for Object-Oriented Programming in Ada 95 The Object Model of Java. Script Implementation of Object-Oriented Constructs 2

Introduction • Many object-oriented programming (OOP) languages – Some support procedural and data-oriented programming (e. g. , Ada and C++) – Some support functional program (e. g. , CLOS) – Newer languages do not support other paradigms but use their imperative structures (e. g. , Java and C#) – Some are pure OOP language (e. g. , Smalltalk) 3

Object-Oriented Programming • Abstract data types • Inheritance – Inheritance is the central theme in OOP and languages that support it • Dynamic Binding – method calls to method body • Polymorphism 4

Inheritance • Productivity increases can come from reuse – ADTs can be reused, but require modifications – Modifications often require changes to all client programs – Person making changes must understand ADT • Inheritance allows new classes defined in terms of existing ones, i. e. , by allowing them to inherit common parts • Inheritance imparts new structure to problem, providing beneficial program organization 5

Object-Oriented Concepts • ADTs are called classes • Class instances are called objects • A class that inherits is a derived class or a subclass • The class from which another class inherits is a parent class or superclass • Subprograms that define operations on objects are called methods 6

Object-Oriented Concepts (continued) • Calls to methods are called messages • The entire collection of methods of an object is called its message protocol or message interface • Messages have two parts--a method name and the destination object • In the simplest case, a class inherits all of the entities of its parent 7

Object-Oriented Concepts (continued) • Inheritance can be complicated by access controls to encapsulated entities – A class can hide entities from its subclasses – A class can hide entities from its clients – A class can also hide entities from its clients while allowing its subclasses to see them • Besides inheriting methods as is, a class can modify an inherited method – The new one overrides the inherited one – The method in the parent is overriden 8

Object-Oriented Concepts (continued) • There are two kinds of variables in a class: – Class variables - one/class – Instance variables - one/object • There are two kinds of methods in a class: • – Class methods – accept messages to the class – Instance methods – accept messages to objects 9

Object-Oriented Concepts (continued) • Single inheritance: a new class is a subclass of a single parent – relationships shown in derivation tree • Multiple Inheritance: a new class has more than one parent class – relationships shown in derivation graph • One disadvantage of inheritance for reuse: – Creates interdependencies among classes which complicate maintenance 10

Dynamic Binding • A polymorphic variable(e. g. , a pointer in C++) can be defined in a class that is able to reference (or point to) objects of the class and objects of its descendants • When a class hierarchy includes classes that override methods and such methods are called through a polymorphic variable, the binding to the correct method may be dynamic • Allows software systems to be more easily extended during both development and maintenance 11

Dynamic Binding Concepts • An abstract method is one that does not include a definition (it only defines a protocol). This is pure virtual in C++. • An abstract class is one that includes at least one abstract method • An abstract class cannot be instantiated 12

Design Issues for OOP Languages • • The Exclusivity of Objects Subclasses as Type Checking and Polymorphism Single and Multiple Inheritance Object Allocation and De-Allocation Dynamic and Static Binding Nested Classes 13

The Exclusivity of Objects • Everything is an object – Advantage - elegance and purity – Disadvantage - slow operations on simple objects • Add objects to a complete typing system – Advantage - fast operations on simple objects – Disadvantage - results in a confusing type system (two kinds of entities) • Include an imperative-style typing system for primitives but make everything else objects – Advantage - fast operations on simple objects and a relatively small typing system – Disadvantage - still some confusion because of the two type systems 14

Are Subclasses Subtypes? • Does an “is-a” relationship hold between a parent class object and an object of the subclass? – If a derived class is-a parent class, then objects of the derived class must behave the same as the parent class object • A derived class is a subtype if it has an is-a relationship with its parent class – Subclass can only add variables and methods and override inherited methods in “compatible” ways 15

Type Checking and Polymorphism • Polymorphism may require dynamic type checking of parameters and the return value – Dynamic type checking is costly and delays error detection • If overriding methods are restricted to having the same parameter types and return type, the checking can be static 16

Single and Multiple Inheritance • Multiple inheritance allows a new class to inherit from two or more classes • Disadvantages of multiple inheritance: – Language and implementation complexity (in part due to name collisions) – Potential inefficiency - dynamic binding costs more with multiple inheritance (but not much) • Advantage: – Sometimes it is extremely convenient and valuable 17

Multiple Inheritance Z display A B display C Which display should C inherit? 18

Allocation and De-Allocation of Objects • From where are objects allocated? – If they behave like the ADTs, they can be allocated from anywhere • Allocated from the run-time stack • Explicitly created on the heap (via new) – If they are all heap-dynamic, references can be uniform thru a pointer or reference variable • Simplifies assignment - dereferencing can be implicit – If objects are stack dynamic, there is a problem with regard to subtypes • Is deallocation explicit or implicit? 19

Allocation and De-Allocation (cont) Class A {. . }; // subclass of A, with an extra field Class B : public A {. . }; // legal statement in main: a 1 = b 1; The information in b 1 will be truncated, could be confusing. a 1 b 1 Not an issue with pointers, which will just copy the address. 20

Dynamic and Static Binding • Should all binding of messages to methods be dynamic? – If none are, you lose the advantages of dynamic binding – If all are, it is inefficient • Allow the user to specify 21

Nested Classes • If a new class is needed by only one class, there is no reason to define so it can be seen by other classes – Can the new class be nested inside the class that uses it? – In some cases, the new class is nested inside a subprogram rather than directly in another class • Other issues: – Which facilities of the nesting class should be visible to the nested class and vice versa 22

Support for OOP in Smalltalk • Smalltalk is a pure OOP language – Everything is an object – All objects have local memory – All computation is through objects sending messages to objects – None of the appearances of imperative languages – All objects are allocated from the heap – All de-allocation is implicit 23

Support for OOP in Smalltalk (cont) • Type Checking and Polymorphism – All binding of messages to methods is dynamic • The process is to search the object to which the message is sent for the method; if not found, search the superclass, etc. up to the system class which has no superclass – The only type checking in Smalltalk is dynamic and the only type error occurs when a message is sent to an object that has no matching method 24

Support for OOP in Smalltalk (cont) • Inheritance – A Smalltalk subclass inherits all of the instance variables, instance methods, and class methods of its superclass – All subclasses are subtypes (nothing can be hidden) – All inheritance is implementation inheritance – No multiple inheritance 25

Support for OOP in Smalltalk (cont) • Evaluation of Smalltalk – The syntax of the language is simple and regular – Good example of power provided by a small language – Slow compared with conventional compiled imperative languages – Dynamic binding allows type errors to go undetected until run time – Greatest impact: advancement of OOP 26

Support for OOP in C++ • General Characteristics: – Evolved from SIMULA 67 – Most widely used OOP language – Mixed typing system (retains C types, adds classes) – Allocation can be static, stack-dynamic or heapdynamic – Constructors and destructors • implicitly called when objects are created/cease to exist – Elaborate access controls to class entities 27

Support for OOP in C++ (continued) • Inheritance – A class need not be the subclass of any class – Access controls for members are – Private (visible only in the class and friends) (disallows subclasses from being subtypes) – Public (visible in subclasses and clients) – Protected (visible in the class and in subclasses, but not clients) 28

Support for OOP in C++ (continued) • In addition, the subclassing process can be declared with access controls (private or public), which define potential changes in access by subclasses – Private derivation - inherited public and protected members are private in the subclasses. Does not represent an is-a relationship. – Public derivation - public and protected members are also public and protected in subclasses 29

Inheritance Example in C++ class base_class { private: int a; float x; protected: int b; float y; public: int c; float z; }; class subclass_1 : public base_class { … }; // In this one, b and y are protected and // c and z are public class // // // subclass_2 : private base_class { … }; In this one, b, y, c, and z are private, and no derived class has access to any member of base_class 30

Reexportation in C++ • A member that is not accessible in a subclass (because of private derivation) can be declared to be visible there using the scope resolution operator (: : ), e. g. , class subclass_3 : private base_class { base_class : : c; … } 31

Reexportation (continued) • One motivation for using private derivation – A class provides members that must be visible, so they are defined to be public members; a derived class adds some new members, but does not want its clients to see the members of the parent class, even though they had to be public in the parent class definition 32

Private derivation example class single_linked_list { private: class node { public: node *link; int contents; } node *head; public: single_linked_list() { head = 0; } void insert_at_head(int); void insert_at_tail(int); int remove_at_head(); bool empty(); }; 33

Private derivation (cont) class stack : public single_linked_list { public: stack() {} void push(int value) { insert_at_head(value); } int pop() { return remove_at_head(); } }; Not only push and pop, but all other functions are available! Alternative: make private derivation, reexport empty. Reasonable because stacks and queues are specializations of linked lists, but not subtypes. 34

Support for OOP in C++ (continued) • Multiple inheritance is supported – If there are two inherited members with the same name, they can both be referenced using the scope resolution operator 35

Support for OOP in C++ (continued) • Dynamic Binding – Methods can be defined to be virtual, which means that they can be called through polymorphic variables and dynamically bound to messages – A pure virtual function has no definition at all – A class that has at least one pure virtual function is an abstract class 36

Dynamic Binding Example class shape { public: virtual void draw() =0; }; class rectangle : public shape { public: void draw() { cout << "rectn"; } }; class square : public rectangle { public: void draw() { cout << "squaren"; } }; 37

Dynamic Binding Example int main() { square *sq = new square; rectangle *rect = new rectangle; shape *ptr_shape = sq; ptr_shape->draw(); // Square rect->draw(); // Rect rect = sq; rect->draw(); // Square sq 2; rectangle r 2 = sq 2; r 2. draw(); // Rect (even though it contains a square) } 38

Support for OOP in C++ (continued) • Evaluation – C++ provides extensive access controls (unlike Smalltalk) – C++ provides multiple inheritance – In C++, the programmer must decide at design time which methods will be statically bound and which must be dynamically bound • Static binding is faster! • Design decision may be wrong, requiring change later • Dynamic binding in C++ is faster than Smalltalk 39

Support for OOP in C++ (continued) • Smalltalk type checking is dynamic (flexible, but somewhat unsafe). Less expensive to catch problems at compile time. • Because of interpretation and dynamic binding, Smalltalk is ~10 times slower than C++ is slightly less efficient than C. • C++ is large and complex. 40

Support for OOP in Java • Because of its close relationship to C++, focus is on the differences from that language • General Characteristics – All data are objects except the primitive types – All primitive types have wrapper classes that store one data value (needed for situations like Array. List container class that must store objects) • my. Arraylist. add(new Integer(10)); – Wrapper classes used less frequently since Java 5. 0 due to automatic coercion called boxing: • my. Arraylist. add(10); 41

Support for OOP in Java (cont) • All classes are derived from Object • All objects are heap-dynamic, are referenced through reference variables, and most are allocated with new • A finalize method is implicitly called when the garbage collector is about to reclaim the storage occupied by the object. Problem: can't predict when this will occur, occasionally causes a problem if need to reclaim resources sooner. 42

Support for OOP in Java (cont) • Inheritance – Single inheritance supported only, but there is an abstract class category that provides some of the benefits of multiple inheritance (interface) – An interface can include only method declarations and named constants, e. g. , public interface Comparable { public int compared. To (Object b); } – A class implements an interface rather than extending it 43

Support for OOP in Java (cont) • Interfaces can be treated as types, so a method can specify an interface as a formal parameter. Any class that implements the interface is allowed. • A variable can also have the type of an interface. Can reference any object of any class that implements the interface. • Interfaces do not provide code reuse. 44

Support for OOP in Java (cont) • Methods can be final (cannot be overriden) • Classes can be final (cannot be extended) 45

Support for OOP in Java (cont) • Dynamic Binding – In Java, all messages are dynamically bound to methods, unless the method is final (i. e. , it cannot be overriden, therefore dynamic binding serves no purpose) – Static binding is also used if the method is static or private both of which disallow overriding 46

Support for OOP in Java (cont) • Several varieties of nested classes • All can be hidden from all classes in their package, except for the nesting class • Nested classes can be anonymous • A local nested class is defined in a method of its nesting class – No access specifier is used – Commonly used for event handlers 47

Support for OOP in Java (cont) • Evaluation – Design decisions to support OOP are similar to C++ – No support for procedural programming – No parentless classes – Dynamic binding is used as “normal” way to bind method calls to method definitions – Uses interfaces to provide a simple form of support for multiple inheritance 48

Support for OOP in C# • General characteristics – – Support for OOP similar to Java Includes both classes and structs Classes are similar to Java’s classes structs are less powerful stack-dynamic constructs 49

Support for OOP in C# (continued) • Inheritance – Uses the syntax of C++ for defining classes – A method inherited from parent class can be replaced in the derived class by marking its definition with new – The parent class version can still be called explicitly with the prefix base: base. Draw() 50

Support for OOP in C# • Dynamic binding – To allow dynamic binding of method calls to methods: • The base class method is marked virtual • The corresponding methods in derived classes are marked override – Abstract methods are marked abstract and must be implemented in all subclasses – All C# classes are ultimately derived from a single root class, Object 51

C# Example public class shape { public: abstract void draw(); }; class rectangle : public shape { public override void draw() { cout << "rectn"; } }; class square : public rectangle { public override void draw() { cout << "squaren"; } }; 52

Support for OOP in C# (continued) • Nested Classes – A C# class that is directly nested in a nesting class behaves like a Java static nested class – C# does not support nested classes that behave like the non-static classes of Java 53

Support for OOP in C# • Evaluation – C# is the most recently designed C-based OO language – The differences between C#’s and Java’s support for OOP are relatively minor 54

Support for OOP in Ada 95 • General Characteristics – OOP was one of the most important extensions to Ada 83 – Encapsulation container is a package that defines a tagged type – A tagged type is one in which every object includes a tag to indicate during execution its type (the tags are internal) – Tagged types can be either private types or records – No constructors or destructors are implicitly called 55

Support for OOP in Ada 95 (continued) • Inheritance – Subclasses can be derived from tagged types – New entities are added to the inherited entities by placing them in a record definition – All subclasses are subtypes – No support for multiple inheritance • A comparable effect can be achieved using generic classes 56

Example of a Tagged Type Package PERSON_PKG is type PERSON is tagged private; procedure DISPLAY(P : in out PERSON); private type PERSON is tagged record NAME : STRING(1. . 30); ADDRESS : STRING(1. . 30); AGE : INTEGER; end record; end PERSON_PKG; with PERSON_PKG; use PERSON_PKG; package STUDENT_PKG is type STUDENT is new PERSON with record GRADE_POINT_AVERAGE : FLOAT; GRADE_LEVEL : INTEGER; end record; procedure DISPLAY (ST: in STUDENT); end STUDENT_PKG; // Note: DISPLAY is being overridden from person_PKG 57

Support for OOP in Ada 95 (continued) • Dynamic Binding – Dynamic binding is done using polymorphic variables called classwide types • For the tagged type PERSON, the classwide type is PERSON’class – Other bindings are static – Any method may be dynamically bound – Purely abstract base types can be defined in Ada 95 by including the reserved word abstract 58

Support for OOP in Ada 95 (continued) • Evaluation – Ada offers complete support for OOP – C++ offers better form of inheritance than Ada – Ada includes no initialization of objects (e. g. , constructors) – Dynamic binding in C-based OOP languages is restricted to pointers and/or references to objects; Ada has no such restriction and is thus more orthogonal 59

Support for OOP in Ruby • General Characteristics – Pure OO language, similar to Smalltalk – Expressions written in infix form, but still evaluated as messages – Class definitions are executable – Current definition of a class is union of all definitions that have been executed – All variables in Ruby are typeless references – All instance data has private access by default which cannot be changed – Must provide getters and setters – Setters have = attached to names 60

Ruby Example class My. Class # A constructor def initialize @one = 1 @two = 2 end #A getter for @one def one @one end #A setter for @one def one=(my_one) @one = my_one end # of class My. Class 61

Support for OOP in Ruby (cont) • Shortcuts are provided for getters/setters – attr_reader : one, : two – attr_writer : one, : two • attr_reader is a function call, precede variable name with : causes name to be used (rather than dereferenced). 62

Support for OOP in Ruby (cont) • Access control in Ruby is dynamic • Default for methods is public 63

Support for OOP in Ruby (cont) • Inheritance – Subclasses are defined using < rather than : – class My. Sub. Class < Base. Class – Method access controls can be changed in subclasses – Multiple inheritance is not supported – Does not support abstract classes or interfaces – Can include modules, so a class effectively inherits the functions defined in that module. Such a module is called a mixin. – Support for dynamic binding is the same as Smalltalk 64

The Object Model of Java. Script • General Characteristics of Java. Script – Little in common with Java • Similar to Java only in that it uses a similar syntax – Dynamic typing (type of a variable can change every time a value is assigned to it) – No classes or inheritance or polymorphism – Variables can reference objects or can directly access primitive values 65

The Object Model of Java. Script • Java. Script objects – An object has a collection of properties which are either data properties or method properties – Appear as hashes, both internally and externally – A list of property/value pairs – Properties can be added or deleted dynamically 66

The Object Model of Java. Script • A bare object can be created with new and a call to the constructor for Object var my_car = new Object(); • References to properties are with dot notation my_car. make = “Ford”; my_car. model = “Contour SVT”; my_car. engine = new Object(); my_car. engine. hp = 200; • Can create simple constructors to be used with new. 67

Java. Script Evaluation • Effective at what it is designed to be – A scripting language – Define dynamic HTML, validate user forms etc. • Inadequate for large scale development • No encapsulation capability of classes – Large programs cannot be effectively organized • No inheritance – Reuse will be very difficult 68

Implementing OO Constructs • Two interesting and challenging parts – Storage structures for instance variables – Dynamic binding of messages to methods 69

Instance Data Storage • Class instance records (CIRs) store the state of an object – Static (built at compile time) • If a class has a parent, the subclass instance variables are added to the parent CIR • Because CIR is static, access to all instance variables is done as it is in records – Efficient 70

Dynamic Binding of Methods Calls • Methods in a class that are statically bound need not be involved in the CIR; methods that will be dynamically bound must have entries in the CIR – Calls to dynamically bound methods can be connected to the corresponding code thru a pointer in the CIR – The storage structure is sometimes called virtual method tables (vtable) – Method calls can be represented as offsets from the beginning of the vtable – Polymorphic variables must simply reference CIR of correct type 71

Dynamic Binding of Methods Calls Single Inheritance public class A { public int a, b; public void draw() {. . . } public int area() {. . } } public class B extends A { public int c, d; public void draw() {. . . } public void sift() {. . . } } code for A's draw code for A's area VTABLE POINTER a b A's vtable code for A's area code for B's draw code for B's sift VTABLE POINTER ac db B's vtable 72

Summary • OO programming involves three fundamental concepts: ADTs, inheritance, dynamic binding • Major design issues: exclusivity of objects, subclasses and subtypes, type checking and polymorphism, single and multiple inheritance, dynamic binding, explicit and implicit de-allocation of objects, and nested classes • Smalltalk is a pure OOL, as is Ruby • C++ has two distinct type system (hybrid) • Java is not a hybrid language like C++; it supports only OO programming • C# is based on C++ and Java • Java. Script is not an OOP language but provides interesting variations 73